Paul E. McKenney (paulmck) wrote,
Paul E. McKenney
paulmck

Confessions of a Recovering Proprietary Programmer, Part XIII

True confession: I was once a serial junk mailer. Not mere email spam, but physical bulk-postage-rate flyers, advertising a series of non-technical conferences. It was of course for a good cause, and one of the most difficult things about that task was convincing that cause's leaders that this flyer was in fact junk mail. They firmly believed that anyone with even a scrap of compassion would of course read the flyer from beginning to end, feeling the full emotional impact of each and every lovingly crafted word. They reluctantly came around to my view, which was that we had at most 300 milliseconds to catch the recipient's attention, that being the amount of time that the recipient might glance at the flyer on its way into the trash. Or at least I think that they came around to my view. All I really know is that they stopped disputing the point.

But junk mail for worthy causes is not the only thing that can be less welcome than its sender might like.

For example, Jim Wasko noticed a sign at a daycare center that read: “If you are late picking up your child and have not called us in advance, we will give him/her an espresso and a puppy. Have a great day.”

Which goes to show that although puppies are cute and lovable, and although their mother no doubt went to a lot of trouble to bring them into this world, they are, just like junk mail, not universally welcome. And this should not be too surprising, given the questions that come to mind when contemplating a free puppy. Has it had its shots? Is it housebroken? Has it learned that furniture is not food? Has it been spayed/neutered? Is it able to eat normal dogfood, or does it still require bottlefeeding? Is it willing to entertain itself for long periods? And, last, but most definitely not least, is it willing to let you sleep through the night?

Nevertheless, people are often surprised and bitterly disappointed when their offers of free puppies are rejected.

Other people are just as surprised and disappointed when their offers of free patches are rejected. After all, they put a lot of work into their patches, and they might even get into trouble if the patch isn't eventually accepted.

But it turns out that patches are a lot like junk mail and puppies. They are greatly valued by those who produce them, but often viewed with great suspicion by the maintainers receiving them. You see, the thought of accepting a free patch also raises questions. Does the patch apply cleanly? Does it build without errors and warnings? Does it run at all? Does it pass regression tests? Has it been tested with the commonly used combination of configuration parameters? Does the patch have good code style? Is the patch maintainable? Does the patch provide a straightforward and robust solution to whatever problem it is trying to solve? In short, will this patch allow the maintainer to sleep through the night?

I am extremely fortunate in that most of the RCU patches that I receive are “good puppies.” However, not everyone is so lucky, and I occasionally hear from patch submitters whose patches were not well received. They often have a long list of reasons why their patches should have been accepted, including:

  1. I put a lot of work into that patch, so it should have been accepted! Unfortunately, hard work on your part does not guarantee a perception of value on the maintainer's part.
  2. The maintainer's job is to accept patches. Maybe not, your maintainer might well be an unpaid volunteer.
  3. But my maintainer is paid to maintain! True, but he is probably not being paid to do your job.
  4. I am not asking him to do my job, but rather his/her job, which is to accept patches! The maintainer's job is not to accept any and all patches, but instead to accept good patches that further the project's mission.
  5. I really don't like your attitude! I put a lot of work into making this be a very good patch! It should have been accepted! Really? Did you make sure it applied cleanly? Did you follow the project's coding conventions? Did you make sure that it passed regression tests? Did you test it on the full set of platforms supported by the project? Does it avoid problems discussed on the project's mailing list? Did you promptly update your patch based on any feedback you might have received? Is your code maintainable? Is your code aligned with the project's development directions? Do you have a good reputation with the community? Do you have a track record of supporting your submissions? In other words, will your patch allow the maintainer to sleep through the night?
  6. But I don't have time to do all that! Then the maintainer doesn't have time to accept your patch. And most especially doesn't have time to deal with all the problems that your patch is likely to cause.

As a recovering proprietary programmer, I can assure you that things work a bit differently in the open-source world, so some adjustment is required. But participation in an open-source project can be very rewarding and worthwhile!
Tags: confessions
Subscribe

  • Post a new comment

    Error

    default userpic

    Your reply will be screened

    Your IP address will be recorded 

    When you submit the form an invisible reCAPTCHA check will be performed.
    You must follow the Privacy Policy and Google Terms of use.
  • 7 comments